Token导航 LogoToken导航TokenDH.com
研究检索敏感数据github未标认证来源可访问许可证需确认审计提醒

debug-prod-issues调试产品问题

Agent Skill

用于围绕 GitHub 仓库、Issue、Pull Request、分支、提交和代码协作流程提供辅助能力。它适合让 Agent 查询项目状态、整理变更、辅助创建或检查协作事项,并把仓库中的信息转成可执行的下一步。使用时需要区分只读查询和写入操作;涉及创建 PR、修改 Issue、推送分支或访问私有仓库时,应确认 token 权限、目标仓库范围和用户授权。

总安装

10,487

周安装

447

GitHub Stars

219

下载量

3,457
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:debug-prod-issues(调试产品问题)
来源仓库:https://github.com/different-ai/agent-bank
仓库路径:skills/debug-prod-issues
安装命令:
npx skills add https://github.com/different-ai/agent-bank --skill 'debug prod issues'
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/different-ai/agent-bank --skill 'debug prod issues'

简介

专用于 Vercel 生产环境问题排查,集成 Neon 数据库检查与函数日志监控。

  • 适用于部署失败、接口超时或数据不一致等线上故障的快速介入。
  • 所有 Vercel 命令需指定 --scope prologe 限定项目范围避免误操作。
  • 必须区分测试环境与生产环境权限,谨慎执行可能影响用户的操作。
  • debug-prod-issues 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

What I Do

  • Debug production issues on Vercel
  • View and filter Vercel function logs
  • Wait for deployments correctly
  • Connect to production Neon database
  • Inspect tables and data for debugging

Vercel Debugging (0 Finance)

IMPORTANT: Always use --scope prologe for the 0 Finance project:

# All Vercel commands need --scope prologe
vercel logs www.0.finance --scope prologe
vercel ls --scope prologe
vercel env ls --scope prologe

Viewing Logs

# Stream live logs (waits for new logs)
vercel logs www.0.finance --scope prologe

# Logs since a specific time
vercel logs www.0.finance --scope prologe --since 5m
vercel logs www.0.finance --scope prologe --since 1h

# Filter logs (pipe to grep)
vercel logs www.0.finance --scope prologe 2>&1 | grep -i "error"
vercel logs www.0.finance --scope prologe 2>&1 | grep "ai-email"

Note: vercel logs can be slow/hang. Use timeout or Ctrl+C if needed:

timeout 30 vercel logs www.0.finance --scope prologe --since 5m 2>&1 | head -100

Checking Deployments

# List recent deployments
vercel ls --scope prologe | head -10

# Check specific deployment status
vercel inspect <deployment-url> --scope prologe

# Get latest deployment URL
vercel ls --scope prologe 2>/dev/null | head -1

Waiting for Deployments

DO NOT just sleep and hope. Use vercel inspect --wait:

# Get the latest deployment URL
LATEST=$(vercel ls --scope prologe 2>/dev/null | head -1)

# Wait for it to be ready (up to 5 minutes)
vercel inspect "$LATEST" --scope prologe --wait --timeout 5m

# Check the status in the output:
#   status: ● Building  -> still building
#   status: ● Ready     -> deployed and live!
#   status: ● Error     -> build failed

Full workflow after pushing code:

# 1. Push your changes
git push origin main

# 2. Wait a few seconds for Vercel to pick it up
sleep 5

# 3. Get the new deployment URL
LATEST=$(vercel ls --scope prologe 2>/dev/null | head -1)
echo "Waiting for: $LATEST"

# 4. Wait for it to complete
vercel inspect "$LATEST" --scope prologe --wait --timeout 5m

# 5. Now your changes are live!

Check deployment details:

# See build info, aliases, and status
vercel inspect <deployment-url> --scope prologe

# See build logs if something failed
vercel inspect <deployment-url> --scope prologe --logs

Triggering a Redeploy

# Redeploy production from current state
vercel --prod --scope prologe

# Or push to git and wait
git push origin main
# Then use the waiting method above

Environment Variables

# List env vars
vercel env ls --scope prologe

# Add an env var (will prompt for value)
echo "value" | vercel env add VAR_NAME production --scope prologe

# Pull env vars to local .env
vercel env pull .env.local --scope prologe

Database Debugging

Environment Setup

CRITICAL: Always load .env.production.local for production database:

import * as dotenv from 'dotenv';
import path from 'path';

// Load production env
dotenv.config({
  path: path.resolve(__dirname, '../.env.production.local'),
});

Or in a script:

cd /path/to/zerofinance/packages/web
pnpm tsx -e "
import * as dotenv from 'dotenv';
dotenv.config({ path: '.env.production.local' });

// Now use db...
import { db } from './src/db';
"

Database Locations

Zero Finance uses Neon Postgres:

EnvironmentHost PatternUsed By
Developmentep-aged-cherry-*Local dev, some scripts
Productionep-wispy-recipe-* or similarVercel deployment, prod data

Always verify which database you're connecting to:

[DB] Connecting to database host: ep-xxxxx-pooler.us-east-1.aws.neon.tech

Key Tables

TablePurpose
user_safesSafe addresses linked to users/workspaces
incoming_depositsIncoming USDC transfers (synced from Safe API)
outgoing_transfersOutgoing transactions from Safes
usersUser records with primaryWorkspaceId
workspace_membersUser-workspace membership
earn_depositsVault deposit records
ai_email_sessionsAI email agent conversation sessions

Script Template

import * as dotenv from 'dotenv';
dotenv.config({ path: '.env.production.local' });

import { db } from './src/db';
import { userSafes, incomingDeposits } from './src/db/schema';
import { eq } from 'drizzle-orm';

async function main() {
  // Your debugging code here
  const safes = await db.select().from(userSafes);
  console.log('Total safes:', safes.length);
}

main()
  .then(() => process.exit(0))
  .catch(console.error);

Common Issues

1. "Internal server error" from API endpoint

Debug steps:

  1. Check Vercel logs for the specific endpoint
  2. Look for stack traces or error messages
  3. Test with curl to see response: curl -s -X POST "https://www.0.finance/api/endpoint" \ -H "Content-Type: application/json" \ -d '{"test": true}'

2. Code changes not taking effect

Causes:

  • Deployment not complete yet
  • Environment variables not updated (need redeploy)
  • Caching issues

Fix:

  1. Verify new deployment exists: vercel ls --scope prologe | head -3
  2. Check deployment time matches your push
  3. Force redeploy if needed: vercel --prod --scope prologe

3. Safe Not Found for Workspace

Debug:

const safe = await db.query.userSafes.findFirst({
  where: eq(userSafes.safeAddress, '0x...'),
});
const user = await db.query.users.findFirst({
  where: eq(users.privyDid, safe.userDid),
});
console.log('Safe workspace:', safe.workspaceId);
console.log('User primary workspace:', user.primaryWorkspaceId);

4. Environment variable not working

# Check it's set in Vercel
vercel env ls --scope prologe | grep VAR_NAME

# Check which environments it's set for (production, preview, development)
# May need to redeploy for changes to take effect

When to Use This Skill

  • Investigating production API errors
  • Checking if deployments completed
  • Viewing function logs
  • Debugging database/data issues
  • Verifying environment variables
  • Running one-off database scripts

Integration with Testing Strategy

This skill is part of the testing pyramid. Use it when:

  1. Staging tests pass but production fails → Check prod logs
  2. Need to verify fix in production → Inspect DB state
  3. Feature works locally but not in prod → Compare environments

Related Skills

ScenarioSkill to Load
Testing on staging firsttest-staging-branch
Making code testabletestability
After debugging, capture learningsskill-reinforcement

Debugging Workflow (Fast → Slow)

1. Check Vercel logs first (fastest)
2. If unclear, inspect production DB
3. If still unclear, reproduce locally
4. After fix, test on staging before prod
5. Update this skill with new patterns

Learnings Log

Append new discoveries here

2026-01-13: prod DB script must load env before db import

Symptom: POSTGRES_URL missing or wrong host while using .env.production.local. Root Cause: packages/web/src/db/index.ts loads .env.local on import, which can run before dotenv config. Fix: In one-off scripts, call dotenv.config({path: '.env.production.local'}) first and dynamically import ./src/db afterward. Prevention: Avoid static imports of ./src/db in CLI scripts; use dynamic import after dotenv setup.

Template for New Learnings

### YYYY-MM-DD: [Issue Description]

**Symptom**: [What you saw]
**Root Cause**: [Why it happened]
**Fix**: [How to resolve]
**Prevention**: [How to avoid in future]

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

34.54%
按下载量换算1,194

Claude

32.31%
按下载量换算1,117

Cursor

20.53%
按下载量换算710

Gemini CLI

9.22%
按下载量换算319

安全审计

Gen Agent Trust Hub

通过

Socket

可疑

Snyk

通过

权限和风险

敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills